AQUA

AQUA (Automated Quality Assurance) is a tool developed by Cambridge Cognition for automated quality assurance of CNS clinical scales. It analyses audio recordings of patient consultations to automatically detect deviations in scale administration and scoring, with the goal of reducing errors that can affect clinical trial outcomes. AQUA is intended for use in CNS drug development trials, where scale administration errors represent a significant risk to data quality and trial success.

Clinical trials in the CNS space face high failure rates, with 85% of Phase 2 and 3 trials failing. Contributing factors include insensitive scales and error-prone administration. Traditional manual quality control is slow, resource-intensive, and difficult to scale. AQUA addresses this by automating the quality review process, with costs reported at up to 60% lower than traditional providers (based on comparison with similar services).

Key Capabilities

• Automated analysis of audio recordings captured during CNS clinical scale administration
• Detection of deviations in how prompts are read, the sequence in which they are delivered, and whether appropriate follow-ups were made
• Speaker identification and transcription, with labelling of subsections within a scale (e.g., "Word recall" and "Commands" for ADAS-Cog)
• Generation of per-scale quality reports specifying detected quality indicators and their locations within the assessment
• Reviewable transcripts and audio available alongside quality reports
• Customisable quality indicators according to sponsor study requirements

Supported Clinical Scales

• ADAS-Cog
• HAM-A
• MADRS
• CDR
• HAM-D
• PANSS
• Additional scales available at sponsor request

Supported Languages

• English
• Spanish
• French
• Italian
• German
• Dutch
• Polish
• Arabic

Workflow Steps

1. Initiate recording: The rater starts an audio recording at the beginning of the session, capturing all voices including the participant, rater, and any additional speakers.
2. Transcribe and label: Speakers are identified and transcribed, and subsections within the scale are marked for structured analysis.
3. Quality check: The transcript and audio are scanned for scale-specific quality indicators, verifying that prompts were read correctly, in the correct sequence, and that appropriate follow-ups were conducted.
4. Generate report: A quality report is produced for each scale, detailing detected quality indicators and their locations within the assessment.

Deployment and Integration

• Powered by Winterlight Labs' speech analysis platform
• Compatible with Cambridge Cognition's eCOA platform or third-party eCOA providers
• Designed to scale across trials of varying sizes
• Cambridge Cognition partners with ratings training providers and CROs to support timely remediation of identified issues